Interaction between tramadol and phenprocoumon

Abstract

We report an interaction between phenprocoumon (Marcoumar) and tramadol which led to an increase in International Normalised Ratio (INR) in two patients. A 54-year-old woman, otherwise healthy, with a suspected deep venous thrombosis in the left lower leg, was begun on heparin (10 000 IU twice daily) and phenprocoumon monitored by INR (therapeutic target range 2–3). Treatment was stopped after 2 days, because the diagnosis of deep venous thrombosis was not confirmed by ultrasonography. Tramadol (Nobligan, Grunenthal, Grunenthal GmbH, Aachen, Germany) 50–100 mg four times daily was given from day 3 for low back pain (figure). Her only other medication was paracetamol 1 g four times daily. As the INR kept rising despite withdrawal of phenprocoumon, phytomenadione (vitamin K1) 2 mg (day 5) and 5 mg (days 9–11) was given orally and the INR gradually fell to 0·9 (figure). On day 6, she had a nosebleed (INR 4·0). A 66-year-old woman with idiopathic lung fibrosis was admitted after a fall that resulted in fracture of a lumbar vertebra. Tramadol (Dolol, Nocomed Christianes, Chausee de Gand, Brussels, Belgium) 100 mg four times daily was prescribed for pain. Her other medication included estazolam 250 mg four times daily, diflunisal 50 mg three times daily, oxazepam 30 mg daily, calcium (400 mg) and vitamin D (5 g) twice daily, and paracetamol 1 g four times daily. 1 month later, she had a pulmonary embolism after an endoscopic retrograde cholangiopancreaticography, and treatment with heparin (12 500 IU twice daily) and phenprocoumon was begun. A rise in INR was seen and on day 6 phytomenadione 2 mg was given intramuscularly. Tramadol was stopped on day 8 and 3 days later, INR was stable at about 2 (figure). Albumin concentration, prothrombin time, and plasma transaminases were normal in both patients. Apart from tramadol, none of the prescribed drugs was suspected of affecting the anticoagulant effect of phenprocoumon. A pharmacodynamic interaction in which tramadol potentiated the action of phenprocoumon on vitamin K epoxide reductase may have caused the prolonged rise in INR. Several pharmacokinetic mechanisms are also possible. Increased absorption of phenprocoumon is unlikely because phenprocoumon is almost completely absorbed. An interaction caused by protein displacement of phenprocoumon (plasma protein binding 98–99%) and tramadol (plasma protein binding 20%) is unlikely because of the low protein binding of tramadol. A change in apparent volume of distribution is unlikely too, as none of the patients were oedematous at any time and both had normal serum creatinine. Tramadol is metabolised by CYP2D6, and phenprocoumon is probably metabolised by CYP2C9 but even so, we think the most likely mechanism for this interaction is inhibition of phenprocoumon metabolism in the liver by tramadol or one of its metabolites, as described for the drug interaction between warfarin and phenylbutazone. If tramadol is widely prescribed to older people taking anticoagulants this interaction should be monitored by frequent INR measurements. This advice also applies to warfarin, because adverse reports to the Danish Medicines Agency’s Committee on Adverse Drug Reactions suggest that tramadol increases the effect of both warfarin (three reports) and phenprocoumon (two reports).